evdwl goes to zero after 1st step in reaxff simulation

Hi all,

I am doing a trial Ni NVT Reaxff simulation using lammps. The evdwl term in *.pot file goes to zero after the 1st step. This gives energy discrepancies between stand alone reaxff code and lammps implementation. I am using 28March-2012 version. Can you please tell me if this is a known error?

My input file:

units real
atom_style charge
boundary p p p

read_data Ni_900k_bulk.data

thermo_style custom etotal pe ke vol press temp pxx pyy pzz pxy pxz pyz lx ly lz xy xz yz
thermo_modify line multi flush yes
thermo 100

neighbor 2 bin
neigh_modify every 10 delay 0 check no
pair_style reax/c lmp_control_900k_bulk checkqeq no
pair_coeff * * ffield.reax 6

dump 1 all custom 1000 Ni_900k_bulk_nvt.dump id type xu yu zu
timestep 1.0
velocity all create 900.0 4928459 rot yes dist gaussian
fix 1 all nvt temp 900.0 900.0 100.0
run 30000
write_restart Ni_900k_bulk_equil_nvt.rst

The first lines of out file:

---------------- Step 0 ----- CPU = 0.0000 (sec) ----------------
TotEng = -202005.5731 PotEng = -207497.1194 KinEng = 5491.5464
Volume = 23200.0904 Press = -45566.3910 Temp = 900.0000
Pxx = -45353.6837 Pyy = -45474.7288 Pzz = -45870.7606
Pxy = 102.8131 Pxz = -237.9743 Pyz = 321.6566
Lx = 28.5209 Ly = 28.5209 Lz = 28.5209
Xy = 0.0000 Xz = 0.0000 Yz = 0.0000
---------------- Step 100 ----- CPU = 87.5354 (sec) ----------------
TotEng = -276456.3085 PotEng = -279660.9110 KinEng = 3204.6025
Volume = 23200.0904 Press = -27807.2680 Temp = 525.1967
Pxx = -27671.1583 Pyy = -27955.9491 Pzz = -27794.6966
Pxy = 650.9254 Pxz = 136.7142 Pyz = 62.0146
Lx = 28.5209 Ly = 28.5209 Lz = 28.5209
Xy = 0.0000 Xz = 0.0000 Yz = 0.0000

Corresponding lines from the *.pot file generated by lammps
step ebond eatom elp eang ecoa ehb etor econj evdw ecoul epol
0 -611258.41 331621.80 0.00 0.00 0.00 0.00 0.00 0.00 72139.50 0.00 0.00
100 -610570.72 330909.81 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

regards,
Karthik Guda
Penn State University

Hi Karthik,

First, update to the most recent distribution and do not use any
control file. The default control settings have been changed to the
same as that of the stand-alone code.

That is actually step 100 in your log file, instead of 1st step. Can
you compare all itemized energies from LAMMPS and the stand-alone
code?

Ray

Hi,

Surprisingly, the problem was solved setting the control file to NULL. The evdwl is no longer zero and also “compute 1 all pair reax” didn’t work but “compute 1 all pair reax/c” worked. Should the online documentation be updated?

regards,
Karthik

Hi Karthik,

Thanks, good to know you have resolved the issue.

I believe the default control settings on the doc page is up to date
and correct. Also, "compute 1 all pair reax/c" is to be used with
"pair_style reax/c", and "compute 1 all pair reax" with "pair_style
reax". SInce you used pair_style reax/c, "compute pair reax"
certainly does not work.

Best,
Ray